Artificial Intelligence (AI) is changing the way real estate professionals approach investment analysis, financial modelling and due diligence. Tasks that once required hours of reviewing documents, extracting information and manually updating financial models can now be completed much faster with the help of AI tools.
But how much can we really rely on AI when making investment decisions? And how can real estate analysts use it effectively without compromising accuracy, professional judgement or data security?
These questions were explored in a recent webinar hosted by Cambridge Finance, in collaboration with The Real Estate & Private Equity Analyst Network, featuring Maria Wiedner, Founder and CEO of Cambridge Finance, and Jason Salmon, AI technology trainer and consultant.
The session explored practical ways to incorporate AI into real estate investment analysis, from improving prompts and updating Excel financial models to analysing large document sets, building due diligence dashboards and understanding the future of AI agents.
Here are the key insights from the discussion.
1. Understanding Where AI Adds Value in Real Estate Investment Analysis
One of the most important starting points when adopting AI is understanding which tasks genuinely benefit from it.
According to Maria Wiedner, the best opportunities often come from identifying repetitive, time-consuming activities that do not require extensive professional judgement but still take up a significant amount of an analyst’s working day.
In real estate investment analysis, these tasks might include:
- Reviewing and summarising investment brochures.
- Extracting key information from tenancy schedules.
- Comparing rental information across multiple documents.
- Preparing initial investment memorandums.
- Updating existing financial models with new property information.
- Organising due diligence findings.
- Preparing reports and summaries for investment committees.
AI can help automate much of the information-processing work involved in these activities, allowing analysts to spend more time interpreting results, assessing risks and making informed recommendations.
However, AI should be viewed as a tool that supports the analyst rather than replaces their expertise.
The fundamental requirement remains the same: professionals must understand the investment, the underlying assumptions and the calculations before relying on any output produced by AI.
2. Better Prompting: Getting More Accurate and Useful Results
The quality of an AI-generated response depends significantly on the instructions provided.
Simply uploading a property brochure and asking an AI tool to analyse it may produce a useful summary, but the results can be inconsistent or lack the specific information required for an investment decision.
Maria introduced the RTCFC prompting framework, which provides a structured approach to creating more effective prompts.
The framework covers five essential elements:
Role: Define who the AI should act as. For example, a senior investment analyst at a UK real estate fund.
Task: Clearly explain what you want the AI to accomplish. This could involve summarising the key investment risks in an acquisition brochure.
Context: Provide relevant background information about the asset or transaction, such as the property type, location, investment strategy and lease structure.
Format: Specify how the information should be presented. For example, as an investment committee memorandum, risk summary or due diligence dashboard.
Constraints: Set clear requirements, such as using UK English, referencing source document page numbers and highlighting any information that cannot be verified.
This structured approach gives the AI a much clearer understanding of the expected output.
For real estate professionals, it also helps ensure that the results are relevant to the actual investment task rather than simply producing a general summary.
3. Turning Property Brochures into Investment Analysis
One of the most practical demonstrations during the webinar involved analysing a real estate investment portfolio using AI.
Maria used a 45-page acquisition brochure for a reversionary multi-let logistics portfolio, referred to as the Nexus Portfolio.
The portfolio included information about individual properties, occupancy, tenancy schedules, rental income and the proposed acquisition price.
The task was to assess whether the portfolio was worth acquiring at the proposed price of £64 million.
Rather than manually reviewing all 45 pages and extracting the relevant information, Maria demonstrated how an AI tool could process the brochure and prepare an initial investment committee memorandum.
Using a structured prompt, the AI generated:
- An investment summary.
- Key financial information.
- Headline and passing rental information.
- Net initial yield and reversionary yield analysis.
- Portfolio and occupancy details.
- Potential investment risks.
- An initial acquisition recommendation.
One particularly interesting finding was the difference between the headline net initial yield of approximately 6% and the passing rent yield of approximately 4.42%.
This highlighted an important point: headline figures do not always tell the complete investment story.
The analyst still needs to understand what sits behind the numbers, including rent top-ups, void guarantees and the underlying rental assumptions.
The AI also referenced the relevant pages of the original brochure, allowing the analyst to trace information back to its source.
This is particularly valuable when reviewing lengthy investment documents, as it provides a starting point for further investigation and verification.
4. Using AI to Update Existing Excel Financial Models
Another significant application demonstrated during the webinar was using AI to update an existing Excel financial model.
Real estate analysts frequently work with established financial modelling templates. When a new investment opportunity arises, they may need to transfer information from acquisition brochures into tenancy schedules, assumptions, dashboards and other sections of their models.
This process can involve considerable copying, pasting and manual data entry.
Maria demonstrated how an existing financial model could be uploaded alongside the Nexus Portfolio brochure, with AI instructed to update the model using the information contained in the acquisition documents.
The AI was able to:
- Update the tenancy schedule.
- Populate relevant property and tenant information.
- Adjust assumptions to reflect the asset type.
- Update the financial model template.
- Provide information about where changes had been made.
An interesting example involved refurbishment costs. The original model had been designed for an office investment, with refurbishment assumptions that were not necessarily appropriate for an industrial property.
The AI adjusted these assumptions to reflect the different asset type.
This demonstrated how AI can help analysts adapt existing models to new investment opportunities.
However, Maria emphasised that the resulting model still requires careful checking.
In particular, analysts must verify important financial calculations, including:
- Internal Rate of Return (IRR).
- Geared IRR.
- Profit on cost.
- Purchase costs.
- Rental assumptions.
- Financing costs and other key inputs.
The most effective approach is not necessarily to ask AI to build an entirely new financial model from scratch.
Instead, analysts can benefit from developing reliable, well-structured financial modelling templates and using AI to update them with new investment information.
This creates greater consistency across transactions and reduces the risk of ending up with different modelling structures for every deal.
5. Preparing Investment Committee Memorandums with AI
The same approach can also be applied to investment committee reporting.
Many organisations have established investment memorandum templates that require analysts to present information in a consistent format.
Maria demonstrated how an existing investment committee template could be uploaded alongside the property brochure, allowing AI to populate the document with relevant investment information.
The AI could also use online research to supplement the initial information and help produce a more developed investment memorandum.
This can significantly reduce the time spent preparing initial reports.
However, the analyst must still verify the information, particularly where the memorandum includes financial assumptions, acquisition costs, market information or investment recommendations.
The objective is to accelerate the preparation process while maintaining the quality and consistency expected in professional investment reporting.
6. Retrieval-Augmented Generation (RAG): Searching Large Volumes of Real Estate Documents
While individual document analysis is useful, real estate transactions often involve much larger volumes of information.
Due diligence exercises, for example, can involve hundreds of documents, including leases, loan agreements, valuation reports, survey reports and other supporting materials.
Searching through these documents manually can be extremely time-consuming.
Jason Salmon introduced Retrieval-Augmented Generation (RAG), a technology that allows AI to search an organisation’s own documents and use the relevant information to answer questions.
Rather than simply relying on the general knowledge of an AI model, RAG connects the model to a specific collection of information.
Jason explained that the process involves two main stages.
Stage 1: Indexing
The documents are processed and broken down into smaller sections, known as chunks.
These sections are then converted into numerical representations called embeddings, which help the system identify relationships between different pieces of information.
The resulting information is stored in a searchable database.
Stage 2: Retrieval and Generation
When a user asks a question, the system searches the indexed information to identify the most relevant sections.
That information is then provided to the AI model, which uses it to generate an answer based on the organisation’s own documents.
For example, an analyst could ask:
- Which leases have a tenant break option before 2028?
- Which leases are approaching expiry?
- What is our exposure to upcoming lease expirations?
- Are there any service charge costs above budget?
- What are the key obligations contained in the loan agreements?
Instead of manually opening multiple documents, users can ask questions in plain English and receive answers based on the available information.
7. Practical Applications of RAG in Real Estate
Jason highlighted several areas where RAG could be particularly useful across real estate investment and finance.
Due diligence: Reviewing large collections of documents and identifying potential risks or inconsistencies.
Lease abstraction: Extracting important lease information, including break clauses, expiry dates and tenant obligations.
Loan and facility agreements: Identifying relevant terms and conditions across financing documents.
Valuation and survey reports: Searching reports for specific assumptions, findings and relevant information.
Funding and investor reporting: Retrieving information from different documents to support reporting and analysis.
Internal knowledge management: Creating a searchable knowledge base that allows employees to find information without knowing exactly where individual documents are stored.
One of the biggest advantages is accessibility. Employees do not necessarily need specialist technical knowledge to retrieve information. They can ask questions in everyday language and receive relevant responses.
As with other AI applications, the system should provide references to its information sources so that users can verify the answers.
8. AI-Powered Due Diligence Dashboards: Making Financial Information Easier to Understand
Another important area discussed during the webinar was the use of AI to improve the presentation and interpretation of financial information.
Financial analysts are accustomed to working with spreadsheets, detailed calculations and large amounts of numerical data.
However, senior management and other decision-makers may prefer information presented visually.
Jason demonstrated how AI could be used to turn property financial information into an interactive due diligence dashboard.
Using a small number of prompts and the underlying property information, the AI generated a visual tool that presented key findings in a more accessible format.
The dashboard could help users:
- Compare financial figures against information presented in the acquisition brochure.
- Identify inconsistencies in the underlying data.
- Highlight potential issues and risks.
- Review important property information visually.
- Understand key findings without navigating multiple spreadsheets.
This is particularly useful when presenting investment information to senior leadership teams and investment committees.
Rather than replacing detailed financial models, these dashboards provide another way to communicate the information contained within them.
9. Data Security, Accuracy and the Importance of Data Quality
While AI offers significant opportunities, the webinar also highlighted some important challenges.
Maria stressed that AI can generate incorrect information, make assumptions or present figures that appear reasonable but cannot be supported by the underlying evidence.
The quality of the output depends heavily on the quality of the information provided.
As the speakers emphasised, poor-quality data will produce poor-quality results.
For real estate organisations, this means ensuring that information relating to tenants, investments, financial performance and historical transactions is accurate and properly organised.
Jason also discussed the importance of data security and data sovereignty.
Organisations must consider where their information is stored, who can access it and how it is processed.
There are different approaches to implementing AI systems.
Using external AI services: Organisations can use established platforms and their computing infrastructure to process information. This can offer speed and convenience, but businesses must assess the relevant data protection and security arrangements.
Using locally hosted AI models: Organisations can run open-source AI models within their own infrastructure, giving them greater control over their data. However, this can require additional hardware, technical expertise and investment, and performance may be slower.
Using a hybrid approach: Businesses can combine external AI services with locally managed data systems, depending on their security requirements and the type of work being undertaken.
Jason explained that smaller organisations do not necessarily need to invest heavily in their own hardware to benefit from AI.
Tools such as NotebookLM, Make.com and n8n can provide accessible ways to explore document-based AI and automate smaller workflows.
The key is to identify where AI provides genuine value while maintaining appropriate control over information.
10. Understanding AI Agents and Automated Workflows
The webinar also explored the growing role of AI agents.
Unlike a traditional AI chatbot, which generally responds to individual questions, an AI agent can be given an objective and determine the steps required to achieve it.
Jason explained this using a simple example: making a cup of tea.
An agent needs to understand the sequence of actions involved, identify the tools required and complete each step in the correct order.
The same principle can be applied to business processes.
For example, an AI agent could potentially:
- Retrieve relevant property documents.
- Extract the required financial information.
- Compare the information against an existing financial model.
- Identify missing or inconsistent data.
- Prepare an initial report.
- Present the findings for human review.
The important difference is that the AI is not simply generating a response. It is planning and carrying out a series of tasks using available tools.
This creates opportunities to automate more complex workflows across real estate investment and financial analysis.
11. What Is Model Context Protocol (MCP)?
Jason also introduced Model Context Protocol (MCP), which helps AI models connect with external tools and services.
In simple terms, MCP provides a standardised way for AI systems to understand what different tools can do and how to interact with them.
For example, an AI agent could connect to:
- Internal company databases.
- File systems and document repositories.
- Accounting software.
- External APIs.
- Reporting and presentation tools.
This means an AI system can potentially retrieve information from one source, process it and then use another tool to produce a report or presentation.
The combination of AI agents and MCP could make it easier for organisations to connect previously separate systems and automate workflows across different departments.
12. The Future of AI in Real Estate: Will We Need Fewer Analysts?
One of the most interesting discussions during the webinar focused on the impact of AI on employment and the future role of real estate analysts.
As AI becomes increasingly capable of completing financial modelling tasks, document reviews and investment analysis more quickly, an important question arises: will organisations need fewer analysts?
Maria suggested that the role of the analyst is likely to evolve.
Historically, analysts may have focused heavily on Excel and financial modelling. However, future professionals will need a broader understanding of the investment process.
This includes:
- Financial modelling and Excel.
- Property valuation.
- Construction and development.
- Real estate markets.
- Investment analysis.
- Critical thinking and commercial judgement.
- The ability to understand and verify AI-generated outputs.
Jason also highlighted the importance of understanding the fundamentals.
Simply asking AI to complete a financial model does not mean the analyst understands how the model works.
During the discussion, the speakers noted that when candidates submit AI-assisted modelling tests, employers still need to assess whether they understand and can explain the work they have produced.
The ability to explain the assumptions, calculations and investment conclusions remains essential.
AI may reduce the time spent on repetitive tasks, but professional knowledge and judgement will continue to be important.
13. How Should Real Estate Professionals Start Using AI?
For professionals who are still developing their understanding of AI, Maria recommended starting with small, practical applications.
Rather than trying to automate an entire investment process immediately, identify one repetitive task that takes considerable time.
For example:
- Start by summarising one investment brochure.
- Use AI to extract information from a tenancy schedule.
- Upload an existing financial model and test whether AI can update selected inputs.
- Create a structured investment memorandum.
- Experiment with searching a collection of documents.
Once the results have been tested and verified, the process can be developed further.
The most important point is to understand what the AI is doing and why it has produced a particular result.
There is no shortcut for developing investment knowledge. AI can help accelerate research and analysis, but professionals still need to investigate unfamiliar asset classes, understand different markets and consult relevant experts.
A reliable financial model should remain consistent, structured and appropriate for the specific asset class and geographical market being analysed.
Conclusion: AI Should Strengthen Investment Analysis, Not Replace Professional Judgement
AI is creating new opportunities for real estate professionals to work more efficiently, particularly in areas involving repetitive data processing, document analysis, financial modelling and reporting.
From analysing 45-page acquisition brochures to updating Excel models, searching hundreds of documents and building interactive due diligence dashboards, the webinar demonstrated how these tools can be applied to real investment scenarios.
However, the central message from Maria Wiedner and Jason Salmon was clear: AI-generated information must always be understood, checked and verified.
The strongest analysts will not simply be those who know how to use AI tools. They will be those who combine AI capabilities with sound financial modelling knowledge, property expertise and critical thinking.
The future of real estate investment analysis is therefore not just about doing things faster. It is about using technology to improve the quality of analysis, make better use of information and support more informed investment decisions.
Learn More About AI Applications in Real Estate
Cambridge Finance provides practical training designed to help real estate professionals understand and apply AI in their day-to-day work.
Its AI Applications in Real Estate course explores practical applications of AI, financial modelling, document analysis, RAG systems and AI agents through hands-on exercises.
Visit the Cambridge Finance website to learn more about upcoming courses and training opportunities.
